Between Phonology and Typology. Consonant Duration in Two Gallo-Italian Dialects1
Total Page:16
File Type:pdf, Size:1020Kb
DOI: 10.17469/O2106AISV000016 LORENZO FILIPPONIO, DAVIDE GARASSINO, DALILA DIPINO Between phonology and typology. Consonant duration in two Gallo-Italian dialects1 Segmental correspondences are sometimes due to completely different factors. We exempli- fy this claim by considering the allophonic gemination of post-stress consonants in Bologna and Porto Maurizio dialects, two varieties which display contrastive vowel length. After the experimental confirmation of vowel length contrasts and the existence of differences in the duration of the post-stress consonants, we qualitatively analyze intensity contours. Despite similar duration values, in light of different intensity patterns and rhythm accounts, we hypothesize that post-stress gemination derives from close contact in Bolognese, due to a stronger compensative pattern, while it is residual or due to the pressure of standard Italian in Portorino. Key words: Consonant duration, Northern Italo-Romance, Dialectology, Phonology, Typology. Introduction The aim of this paper is to propose a first experiment dealing with the coexistence of vowel length oppositions and a longer duration of post-stress consonants after short stressed vowels in two Northern Italo-Romance dialects. We will do this in § 4 by measuring the quantity of stressed vowels and post-stress consonants (the meth- ods are explained in § 3). A further, particularly relevant aim is trying to find out whether the correspondence of these features must be ascribed to a similar rhythm pattern or not, which will lead us to some typological considerations about rhythm. In order to do that, we will provide a short introduction to intensity contours (§ 5) before concluding with a brief discussion (§ 6). Before starting our analysis, we will explain what is meant here by rhythm (§ 1), since we will consider single (phono- logical) words and not longer speech chains, and we will give a brief description of length patterns in Gallo-Italian Dialects (§ 2). 1. Some preliminary remarks on rhythm We call rhythm the distribution of features like stress, quantity and tone (the su- prasegmentals analyzed by Lehiste, 1970) among the segments inside a given unit. 1 The paper has been jointly written by the three authors. For academic purposes, LF bears responsi- bility for §§ 1, 2, 6.2; DG for §§ 5, 6.1; DD for §§ 3, 4. While being solely responsible for any weak- nesses or inaccuracies that may be found here, the authors would like to thank Chiara Celata, Franco Cutugno, Christine Mooshammer and an anonymous reviewer for their helpful comments. 270 LORENZO FILIPPONIO, DAVIDE GARASSINO, DALILA DIPINO Since tone does not play a relevant role in the varieties we are going to deal with, we will refer here only to stress and quantity as rhythm-sensitive features. To classify the different typologies of rhythm, we do not adopt the classical categories of stress isochrony and syllable isochrony (Abercrombie, 1967 after Pike, 1945, and thereaf- ter a large amount of studies). We prefer instead those of controll and compensation (first in Vékás, Bertinetto, 1991, then in Bertinetto, Bertini, 2008; see Filipponio, 2012a), which are explainable as the tendencies either to keep the distribution of quantity stable or to unbalance it, normally in favour of the prominent elements (viz. the stressed syllables). In this way, segmental changes leading to a bigger imbal- ance between more and less prominent elements can be ascribed to a compensating pattern, while changes balancing them (or at least no changes) may be due to a con- trolling pattern. Diachronically, these patterns are cyclical (Filipponio, 2012a), so that a controlling pattern can involve a rhythm structure previously modified by a compensating one; at the end, a total compensation should produce an alignment of prominences building a new control phase pattern. The question is whether it is possible to deal with rhythm by describing process- es involving a single (phonological) word – which, in other words, means asking whether a phonological word can be considered as a rhythm unit. Since phono- logical words are normally realized inside the speech chain, viz. an environment dominated by inter- and intra-speaker (context-sensitive) variation as well as a huge amount of coarticulation (assimilation, elision and so forth), the attempt to obtain a rhythm picture of a language starting from this situation should be immediately abandoned. In more general terms, every attempt to extract rhythmic properties from the speech chain seems to be doomed to failure: one can surely agree on this with Arvaniti (2012), who claims that rhythm classifications based on metrics such as %V, ΔV, ΔC (Ramus, Nespor & Mehler, 1999), PVI (Grabe, Low, 2002) and Varco (Dellwo, Wagner, 2003) are unsafe: their unsafety depends exactly on the fact that they try to find out regularities by measuring speech chains (phonetic ut- terances), without regard to prominences (stress, length) and contrasts (stressed/ unstressed, long/short) – what speakers/hearers contrarily do. For all these reasons, the definition of rhythm that we have provided above must be understood as phonological. Therefore, the search for rhythmic features should be exclusively phonology-driven (cf. Dauer, 1983; Bertinetto, 1989; Filipponio, 2012a). In that respect, the phonological word considered as a rhythm unit comes back into play, and the diachrony with it: in the sense that, if some rhythm-driven changes such as lengthening, shortening or loss of segments and/or syllables have been lexicalized, we should be allowed to consider the story of a word structure as a clue to reconstruct the rhythm history of a language. Moreover, allophonic phono- logical rules, as part of the synchronic phonological derivation (Loporcaro, 2015: 234), must be taken into consideration. In sum, we (try to) consider the co-occur- rence of segmental phenomena as epiphenomena related to a superordinate rhythm pattern (see the discussion in § 6). BETWEEN PHONOLOGY AND TYPOLOGY 271 To provide an example of what we mean, consider for instance the Latin propar- oxyton pĕrtĭcă ‘pole’; doing something ‘with a pole’ (instrumental) would have been expressed through the ablative pĕrtĭcā, with a ˈcvc-cv-cvː structure requir- ing a stable control rhythm pattern in order to preserve vowel length contrasts in unstressed syllables (to save the contrast ablative ~ nominative). This pattern was no longer available in the late stages of Latin (cf. Loporcaro, 2015: 10-11; 59) and in early Italo-Romance, as long vowels in unstressed syllables were banned and af- terwards open syllable lengthening came into force as an allophonic phonological rule: all stressed syllables became bimoraic (heavy), so that the Latin variation ˈcv (light), ˈcvː (heavy, open), ˈcvc (heavy, closed), ˈcvːc (superheavy), was reduced to ˈcvː (open), ˈcvc (closed) via ˈcv > ˈcvː and ˈcvːc > ˈcvc. In fact, having only bimoraic stressed syllables and no long vowels in unstressed syllables is already a compensa- tion-oriented pattern, because the moraic weight will always be higher (or at least the same) in stressed syllables than in unstressed ones. Consider now some Italo-Romance outcomes like /ˈpɛrtika/ (Tuscan), /ˈpɛːrdga/ (Eastern Apennine Emilian), and /ˈpɛrtɛgɛ/ (in the Western Lombard tiny village of Monteviasco, see Delucchi, 2016: 170): all in all, they can be interpreted as a result of the influence of different rhythmic patterns. Tuscan has kept (controlled) the late Latin/early Romance pattern (/ˈpɛrtika/ = ˈcvc-cv-cv);2 Emilian has heavily unbalanced the word structure by syncope of the post-stress syllable and the sec- ondary lengthening of stressed vowels (which is regular in the case of a following liquid+plosive cluster: /ˈpɛːrdga/ = ˈcvːcc-cv). In Monteviasco’s dialect, on the contrary, stressed vowel lengthening is absent and unstressed vowels are fixed by full vowel harmony: this weight balance between stressed and unstressed syllable indicates a controlling rhythm pattern (Delucchi, 2016: 310) probably arisen after a compensative phase (Delucchi, Filipponio, 2013)3. Thus, the observation of three different outcomes of the same Latin word brings us to identify three different rhythm patterns. Obviously, all these phonological forms are prone to every kind of phenomenon as soon as they are put in the speech chain. Nevertheless, they do exist in speakers’ knowledge (langue): as said, since rhythm should be understood as a phonological feature, a phonetic analysis dealing 2 With some early compensative oscillation which have left a trace in syncopated proparoxytones (persĭca > pesca ‘peach’, pŏsĭtu(m) > posto ‘posed.part.m.sg, place’ etc., see Rohlfs, 1966: § 138 for further examples), sometimes restored (a control-driven balancing?) with a non-etymological /a/ (jŭvĕne(m) > giovane ‘young.sg’, chrŏnĭca > cronaca ‘chronicle’ etc., see Rohlfs, 1966: § 139 for further examples). 3 The Western Lombard onset-epenthesis pattern /ˈforna/ < *forn < fŭrnu(m), which builds a full disyllabic word, can be considered another evidence of a compensation-to-control path (first apoco- pe, then rebuilding of the unstressed syllable), while the Eastern Emilian rhyme-epenthesis pattern /ˈfawren/ is compatible with the persistent compensation pattern mentioned above – notice also the secondary lengthening of the stressed vowel before liquid+plosive (like in /ˈpɛːrdga/, here /o/ > /aw/), which took place before the apocope (see Repetti, 1995, following Broselow, 1992, for the definition of rhyme-dialects and onset-dialects